Underwater sludge distribution detection method and device, electronic equipment and storage medium

Abstract

The invention provides an underwater sludge distribution detection method and device, electronic equipment and a storage medium, and relates to the technical field of image processing, and the method comprises the steps: obtaining sonar data, and analyzing the sonar data to obtain a sonar waterfall image; performing denoising processing on the sonar waterfall image based on the weight of the first pixel relative to the second pixel in the sonar waterfall image to obtain a target sonar waterfall image; performing super-pixel segmentation on the target sonar waterfall image based on a preset segmentation rule to obtain a super-pixel image; analyzing a local neighborhood mean value and a variance of the super-pixel image to determine a target region super-pixel; determining a target initial contour based on the target area superpixels; and performing level set segmentation processing on the target initial contour to determine a target contour. In the mode, the sonar data is acquired and analyzed, and image denoising, superpixel segmentation and level set segmentation processing are performed to determine the contour of the underwater sludge, so that the contour determination speed and accuracy are ensured.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a method, device, electronic equipment and storage medium for detecting underwater silt distribution. Background Art

[0002] Underwater silt is a general term for silt deposited at the bottom of water bodies, mainly composed of silt soils, which are deposited in still water or slow-flowing water environments, or formed by biochemical reactions. The physical and chemical properties of silt are closely related to its distribution. Accurate detection of underwater silt distribution is of great significance for coastal engineering, water conservancy projects, shipping safety, and ecological and environmental protection. Therefore, how to quickly and accurately determine the contours of underwater silt is an urgent problem to be solved. Summary of the invention

[0003] In view of this, the purpose of the present invention is to provide a method, device, electronic device and storage medium for detecting underwater sludge distribution, which can determine the contour of underwater sludge by acquiring and analyzing sonar data, as well as image denoising, superpixel segmentation and level set segmentation processing, thereby ensuring the speed and accuracy of contour determination.

[0004] In a first aspect, an embodiment of the present invention provides a method for detecting underwater silt distribution, comprising: acquiring sonar data and parsing to obtain a sonar waterfall image; denoising the sonar waterfall image based on the weight of a first pixel relative to a second pixel in the sonar waterfall image to obtain a target sonar waterfall image; performing superpixel segmentation on the target sonar waterfall image based on a preset segmentation rule to obtain a superpixel image; analyzing the local neighborhood mean and variance of the superpixel image to determine the superpixels of the target area; determining the target initial contour based on the superpixels of the target area; and performing level set segmentation on the target initial contour to determine the target contour.

[0005] In a preferred embodiment of the present invention, the above-mentioned denoising the sonar waterfall image based on the weight of the first pixel relative to the second pixel in the sonar waterfall image to obtain the target sonar waterfall image includes: denoising the sonar waterfall image based on the weight of the first pixel relative to the second pixel in the sonar waterfall image to obtain the target sonar waterfall image by the following formula: ;in, ; Indicates the first pixel q i The weight relative to the second pixel p; Z ( p ) represents the weight normalization factor, Represents the second pixel p With the first pixel q i Euclidean distance between centers.

[0006] In a preferred embodiment of the present invention, the target sonar waterfall image is subjected to superpixel segmentation based on a preset segmentation rule to obtain a superpixel image, including: segmenting the target sonar waterfall image according to S×S pixels to obtain multiple superpixel regions; performing pixel seed point iteration and pixel label iteration on the center of each superpixel region; performing center clustering and local superpixel seed point iteration and pixel label iteration within a specified neighborhood of each superpixel region to determine the clustering distance; and allocating each pixel based on the clustering distance to obtain a superpixel image.

[0007] In a preferred embodiment of the present invention, the above-mentioned designated neighborhood is a 2S×2S neighborhood; performing center clustering and performing local superpixel seed point iteration and pixel label iteration in the designated neighborhood of each superpixel area to determine the cluster distance includes: determining the cluster distance by the following formula: ; Where D represents the clustering distance; d c is the pixel distance; d s is the pixel coordinate distance, indicating the coordinate ( x i , y i ) i With coordinates (x j ,y j ) is the position distance of the seed point j; m is a constant; d t Indicates the LDZP texture description value distance.

[0008] In a preferred embodiment of the present invention, the above ;in, d c is the pixel distance, indicating that the pixel value is NL[v] i Pixels i and the pixel value is NL[v] j The difference in pixel values ​​of seed point j; ;in, d s is the pixel coordinate distance, indicating the coordinate (x i ,y i ) i With coordinates (x j ,y j ) at the position distance of the seed point j; ;in, d t is the LDZP texture description value distance, which represents the N-neighborhood LDZP texture description value distance between pixel i and pixel j.

[0009] In a preferred embodiment of the present invention, the silt area in the above-mentioned sonar waterfall image includes a background area, a bright area and a dark area, and the bright area and the dark area are adjacent; the local neighborhood mean and variance of the superpixel image are analyzed to determine the target area superpixel, including: determining a first relationship between the superpixel value of the superpixel in the background area and the local neighborhood mean and variance; determining a second relationship between the superpixel value of the target bright area superpixel in the bright area and the local area mean and variance; determining a third relationship between the superpixel value of the target dark area superpixel in the dark area and the local area mean and variance; determining a local statistical information satisfaction relationship of the target bright area based on the first relationship, the second relationship and the third relationship; determining the target area superpixel based on the local statistical information satisfaction relationship and pre-set data requirements.

[0010] In a preferred embodiment of the present invention, the initial contour of the target includes: an initial contour of a bright area corresponding to the bright area of ​​the target and an initial contour of a dark area corresponding to the dark area of ​​the target; the initial contour of the target is subjected to level set segmentation processing to determine the contour of the target, including: constructing a level set energy function based on distance regularization for the target sonar waterfall image based on the initial contour of the bright area and the initial contour of the dark area; obtaining a dual-region level set evolution equation based on the level set energy function; evolving the initial contour of the bright area and the initial contour of the dark area based on the dual-region level set evolution equation until the evolution converges to obtain the contour of the target.

[0011] In a second aspect, an embodiment of the present invention further provides a detection device for underwater silt distribution, comprising: a sonar data analysis module, used to acquire sonar data and analyze it to obtain a sonar waterfall image; a denoising processing module, used to perform denoising processing on the sonar waterfall image based on the weight of the first pixel relative to the second pixel in the sonar waterfall image to obtain a target sonar waterfall image; a superpixel segmentation module, used to perform superpixel segmentation on the target sonar waterfall image based on a preset segmentation rule to obtain a superpixel image; a target area superpixel determination module, used to analyze the local neighborhood mean and variance of the superpixel image to determine the target area superpixel; a target initial contour determination module, used to determine the target initial contour based on the target area superpixel; a target contour determination module, used to perform level set segmentation processing on the target initial contour to determine the target contour.

[0012] In a third aspect, an embodiment of the present invention further provides an electronic device, including a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the underwater silt distribution detection method of the first aspect above.

[0013] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the underwater silt distribution detection method of the first aspect mentioned above.

[0014] The embodiments of the present invention bring the following beneficial effects: The embodiment of the present invention provides a method, device, electronic device and storage medium for detecting underwater silt distribution, which obtains sonar data and analyzes it to obtain a sonar waterfall image, performs denoising on the sonar waterfall image based on the weight of the first pixel relative to the second pixel in the sonar waterfall image to obtain a target sonar waterfall image, performs superpixel segmentation on the target sonar waterfall image based on a preset segmentation rule to obtain a superpixel image, analyzes the local neighborhood mean and variance of the superpixel image to determine the superpixel of the target area, determines the initial contour of the target based on the superpixel of the target area, and performs level set segmentation on the initial contour of the target to determine the target contour. In this method, the contour of underwater silt is determined by obtaining and analyzing sonar data, as well as image denoising, superpixel segmentation and level set segmentation, thereby ensuring the speed and accuracy of contour determination.

[0015] Other features and advantages of the present disclosure will be set forth in the following description, or some features and advantages may be inferred or unambiguously determined from the description, or may be learned by implementing the above-mentioned technology of the present disclosure.

[0016] In order to make the above-mentioned objectives, features and advantages of the present disclosure more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 A flow chart of a method for detecting underwater silt distribution provided by an embodiment of the present invention; Figure 2 A diagram showing the topography and echo intensity when the sound wave is reflected at the bottom of the water provided by an embodiment of the present invention; Figure 3 A sonar imaging diagram provided by an embodiment of the present invention; Figure 4 A flow chart of another method for detecting underwater silt distribution provided by an embodiment of the present invention; Figure 5 A flow chart of another method for detecting underwater silt distribution provided by an embodiment of the present invention; Figure 6 A schematic diagram of profile evolution provided in an embodiment of the present application; Figure 7 A schematic diagram of the structure of a device for detecting underwater silt distribution provided by an embodiment of the present invention; Figure 8 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] Underwater silt is a general term for silt deposited at the bottom of water bodies, mainly composed of silt soils, which are deposited in still water or slow-flowing water environments, or formed by biochemical reactions. The physical and chemical properties of silt are closely related to its distribution. Accurate detection of underwater silt distribution is of great significance to coastal engineering, water conservancy projects, shipping safety, and ecological and environmental protection.

[0021] Therefore, how to quickly and accurately determine the contour of underwater silt is a problem that needs to be solved urgently.

[0022] Based on this, the embodiments of the present invention provide a method, device, electronic device and storage medium for detecting underwater silt distribution, which can obtain sonar data and analyze it to obtain a sonar waterfall image, perform denoising on the sonar waterfall image based on the weight of the first pixel relative to the second pixel in the sonar waterfall image to obtain a target sonar waterfall image, perform superpixel segmentation on the target sonar waterfall image based on a preset segmentation rule to obtain a superpixel image, analyze the local neighborhood mean and variance of the superpixel image to determine the superpixel of the target area, determine the initial contour of the target based on the superpixel of the target area, and perform level set segmentation on the initial contour of the target to determine the target contour. In this method, the contour of underwater silt is determined by obtaining and analyzing sonar data, as well as image denoising, superpixel segmentation and level set segmentation, thereby ensuring the speed and accuracy of contour determination.

[0023] To facilitate understanding of this embodiment, a method for detecting underwater silt distribution disclosed in an embodiment of the present invention is first introduced in detail.

[0024] Example 1 The embodiment of the present invention provides a method for detecting underwater silt distribution. Figure 1 A flow chart of a method for detecting underwater silt distribution provided by an embodiment of the present invention. Figure 1As shown, the underwater silt distribution detection method may include the following steps: Step S101, acquiring sonar data and analyzing it to obtain a sonar waterfall image.

[0025] Among them, the sonar data can be collected through the sonar equipment, and the sonar data can be saved in the XTF file and analyzed, and the sonar waterfall image can be obtained after the analysis.

[0026] Among them, the sonar waterfall image can be further divided into image frames of fixed size, and the sonar waterfall image can be analyzed frame by frame.

[0027] Specifically, the sonar device amplifies a specific signal through a sonar converter and sends it out in the form of a sound wave. When encountering an obstacle or the bottom of the water, the sound wave is reflected until it returns to the sonar. The size of the echo is received by the converter and the corresponding data is saved. When the water sound wave propagates underwater, the echo intensity is different when it encounters underwater areas with different terrains, and target detection is performed based on the strength of the echo.

[0028] For ease of understanding, Figure 2 The topography and echo intensity mapping diagram when the sound wave is reflected at the bottom of the water provided by the embodiment of the present invention is as follows: Figure 2 As shown in the figure, point ① is the sonar transducer, and the sound waves it emits reach various positions on the bottom of the water in turn. The sound wave echo signal strength corresponding to each position is shown in the envelope below the figure. According to the principle of sonar imaging, the echo signal is strong when the bottom geology is hard, rough or has protrusions. When the bottom geology is soft, flat or concave, the echo signal is weak. In the figure, the terrain at the positions of points ⑤ and ⑥ is protruding, and the echo signal is obviously stronger, while the terrain at the positions of points ④ and ⑩ is concave, and the echo signal is obviously weaker. At the same time, it is noted that at the position of point ⑦, because the sound wave at the position of point ⑦ is blocked by point ⑥ during propagation, the transducer cannot detect the echo signal strength of point ⑦. The same is true for the position of point ⑩. There is no echo signal, which forms a dark area at these two positions in this ton of data. This is an important feature based on sonar for underwater target detection.

[0029] Among them, the sonar equipment can be installed on the underwater robot. For the convenience of understanding, Figure 3The sonar imaging diagram provided by the embodiment of the present invention continuously emits sound wave signals and detects the echo intensity through the transducer during the movement of the underwater robot. Each time a longitudinal sound wave is emitted, the transducer will receive a frame of echo intensity signal as described above, and multiple frames of continuous echo data can form a sonar waterfall image. Continuous scanning of low-intensity data forms continuous low-intensity reflection areas, representing the target area of ​​the terrain protrusion area, while continuous scanning of low-intensity data forms continuous low-intensity reflection areas or blocked areas, representing the terrain depression area or target shadow area. The distance between the target at the bottom of the water and the sonar is greater than the distance between the sonar and the bottom of the water. The target will form an obvious target bright area (white area) outside the bottom line of the water. At the same time, because the rear part of the target area is blocked by the target, an obvious target dark area (black area) will be formed at its corresponding position, and the target bright area is connected to the target dark area.

[0030] Step S102, performing denoising processing on the sonar waterfall image based on the weight of the first pixel relative to the second pixel in the sonar waterfall image to obtain a target sonar waterfall image.

[0031] Among them, when sonar equipment collects sonar data underwater, it is often disturbed by bottom reverberation and the equipment's own noise, resulting in speckle noise mixed in the analyzed sonar waterfall image. Therefore, it is necessary to denoise the sonar waterfall image to obtain the target sonar waterfall image, while smoothing the sonar waterfall image and better retaining the acoustic target characteristics.

[0032] Specifically, performing denoising on the sonar waterfall image based on the weight of the first pixel relative to the second pixel in the sonar waterfall image to obtain the target sonar waterfall image may include: performing denoising on the sonar waterfall image based on the weight of the first pixel relative to the second pixel in the sonar waterfall image to obtain the target sonar waterfall image by the following formula: ;in, ; Indicates the first pixel q i Relative to the second pixel p The weight of; Z (p) represents the weight normalization factor, Represents the second pixel p With the first pixel q i Euclidean distance between centers.

[0033] Among them, for the second pixel in the sonar waterfall image I p , the pixel value after denoising is the weighted sum of each pixel in the rest of the pixel set in the sonar waterfall image, and the second pixel p The neighborhood centered on the first pixel q i The similarity of the neighborhood centered is expressed as w(p , q i ), then w( p , q i ) is the first pixel q i Relative to the second pixel p The weight of the second pixel p The pixel value after denoising is the first pixel q i The size of the contribution made.

[0034] Step S103, performing super-pixel segmentation on the target sonar waterfall image based on a preset segmentation rule to obtain a super-pixel image.

[0035] Among them, the superpixel segmentation method can be used to segment the target sonar waterfall image into superpixel images. Each superpixel value is the pixel average of all original pixels in the superpixel area to obtain a superpixel image, thereby improving the image processing speed and the accuracy of target positioning and shape.

[0036] Specifically, performing superpixel segmentation on the target sonar waterfall image based on a preset segmentation rule to obtain a superpixel image can include: segmenting the target sonar waterfall image according to S×S pixels to obtain multiple superpixel regions; performing pixel seed point iteration and pixel label iteration on the center of each superpixel region; performing center clustering and local superpixel seed point iteration and pixel label iteration within a specified neighborhood of each superpixel region to determine the clustering distance; and allocating each pixel based on the clustering distance to obtain a superpixel image.

[0037] The designated neighborhood is a 2S×2S neighborhood; performing center clustering and performing local superpixel seed point iteration and pixel label iteration in the designated neighborhood of each superpixel region to determine the cluster distance may include: determining the cluster distance by the following formula: ; Where D represents the clustering distance; d c is the pixel distance; d s is the pixel coordinate distance, indicating the coordinate ( x i， y i ) and the pixel point i at the coordinate ( x j , y j ) is the distance between the seed point j and the starting point; m is a constant, which can be 1-40; d t Indicates the LDZP texture description value distance.

[0038] in, ; d c is the pixel distance, indicating that the pixel value is NL[v] i The pixel point i and the pixel value are NL[v] j The difference in pixel value of seed point j.

[0039] in, ; d s is the pixel coordinate distance, indicating the coordinate ( x i , y i ) and the pixel point i at the coordinate ( x j , y j ) is the position distance of the seed point j.

[0040] in, ; d t LDZP texture description value distance, expressed in pixels i With pixels j The distance between the N-neighborhood LDZP texture description values.

[0041] Step S104, analyzing the local neighborhood mean and variance of the superpixel image to determine the target area superpixel.

[0042] Among them, according to the characteristics of underwater target acoustic imaging, the silt area in the target sonar waterfall image can generally be divided into background area, bright area and dark area, and the bright area is adjacent to the dark area. The local neighborhood mean of the target area is the same but in opposite directions, and the local threshold mean of the background area is the same but the variance is small. Therefore, adaptive threshold processing can be used for detection, and the local neighborhood threshold mean and variance of the image are analyzed to determine the superpixel of the target area.

[0043] Step S105, determining an initial contour of the target based on the superpixels in the target area.

[0044] Among them, the initial shape of the target corresponding to the superpixel of the target area in the image after adaptive threshold processing can be used as the initial contour of the target.

[0045] Step S106, performing level set segmentation processing on the initial contour of the target to determine the target contour.

[0046] Among them, a level set energy function based on distance regularization can be constructed for the target sonar waterfall image, and a dual-region level set evolution equation can be obtained based on the level set energy function. The dual-region level set evolution equation is evolved until the evolution converges to obtain the target contour.

[0047] The detection method for underwater silt distribution provided by the embodiment of the present invention can obtain sonar data and analyze it to obtain a sonar waterfall image, perform denoising on the sonar waterfall image based on the weight of the first pixel relative to the second pixel in the sonar waterfall image to obtain a target sonar waterfall image, perform superpixel segmentation on the target sonar waterfall image based on a preset segmentation rule to obtain a superpixel image, analyze the local neighborhood mean and variance of the superpixel image to determine the target area superpixel, determine the target initial contour based on the target area superpixel, and perform level set segmentation on the target initial contour to determine the target contour. In this method, the speed and accuracy of contour determination are guaranteed by obtaining and analyzing sonar data and performing image denoising, superpixel segmentation and level set segmentation to achieve the determination of the contour of underwater silt.

[0048] Example 2 The embodiment of the present invention also provides another method for detecting underwater silt distribution; this method is implemented on the basis of the method in the above embodiment; this method focuses on describing the specific implementation method of analyzing the local neighborhood mean and variance of the superpixel image to determine the superpixel of the target area.

[0049] Figure 4 A flow chart of another method for detecting underwater silt distribution provided by an embodiment of the present invention is as follows: Figure 4 As shown, the silt area in the sonar waterfall image includes a background area, a bright area and a dark area, and the bright area and the dark area are adjacent. The local neighborhood mean and variance of the superpixel image are analyzed to determine the superpixel of the target area, which may include the following steps: Step S201, determining a first relationship between a superpixel value of a superpixel in a background area and a local neighborhood mean and variance.

[0050] Among them, for the background area Ω b Super Pixel i , the variance σ of the n×n local neighborhood threshold bi 2 for: ;μ bi 2 Superpixel i The mean of the n×n local neighborhood of the background region is obtained by calculating the superpixel value xbi of the background region superpixel i and the local neighborhood mean μ due to the similar local pixel values ​​in the background region. bi 2 and variance σ bi 2 The first relation is: .

[0051] Step S202, determining a second relationship between the superpixel value of the target bright area superpixel in the bright area and the local area mean and variance.

[0052] Among them, for the target bright area Ω o Superpixel i, μ toi is the n×n local neighborhood mean of superpixel i, x toj is the pixel value of superpixel j in the target bright area. Since the superpixel value in the target bright area is much larger than the superpixel value in the background area, the superpixel in the target bright area can be obtained. i The n×n local neighborhood variance σ bi 2 for: . Then the superpixel in the target bright area i Pixel Value The local area mean and variance σ tbi 2 The second relationship is: .

[0053] Step S203, determining a third relationship between the superpixel value of the target dark area superpixel in the dark area and the local area mean and variance.

[0054] Among them, for the superpixel in the target dark area Ωs i , Superpixel i The n×n local neighborhood mean of The target dark area superpixel j Since the superpixel value in the target dark area is much smaller than the superpixel value in the background area, the superpixel value in the target dark area can be obtained. i The n×n local neighborhood variance for: ; Then the superpixel value xtsi of superpixel i in the target dark area and the local area mean and variance The relationship is: .

[0055] Step S204: determine whether the local statistical information of the target bright area satisfies the relationship based on the first relationship, the second relationship and the third relationship.

[0056] Among them, the local statistical information of the target bright area satisfies the relationship: .

[0057] Step S205, determining the superpixel in the target area based on the local statistical information satisfying the relationship and the pre-set data requirements.

[0058] Specifically, let the local statistical information thresholds μ and σ of the target bright area be 2 Satisfaction relationship: , then the target bright area Ω o Local statistics thresholds μ and σ 2 Satisfaction relationship: , let k0 represent the adaptive processing threshold of the target bright area. When the statistics of superpixel i and its local neighborhood meet the requirements, superpixel i can be considered as the superpixel of the target area: .

[0059] Example 3 The embodiment of the present invention also provides another method for detecting underwater silt distribution; the method is implemented on the basis of the method of the above embodiment; the method focuses on describing the specific implementation method of performing level set segmentation processing on the initial contour of the target to determine the target contour.

[0060] Figure 5 A flow chart of another method for detecting underwater silt distribution provided by an embodiment of the present invention is as follows: Figure 5 As shown, the target initial contour includes: a bright area initial contour corresponding to the target bright area and a dark area initial contour corresponding to the target dark area. The target initial contour is determined by performing level set segmentation processing on the target initial contour, which may include the following steps: Step S301 : constructing a level set energy function based on distance regularization for the target sonar waterfall image based on the initial contour of the bright area and the initial contour of the dark area.

[0061] Specifically, the level set energy function based on distance regularization is constructed for the target sonar waterfall image based on the initial contour of the bright area and the initial contour of the dark area, which can be expressed as the following formula:

[0062] in, is the level set energy function, is the local fitting energy term, is the distance regularization term, is the superpixel segmentation contour constraint.

[0063] Among them, through the target bright area level set function and the target dark region level set function The two zero-order curves divide the sonar waterfall image into three parts: target bright area, target dark area and background area, and construct the local fitting energy of the three parts according to the pixel values. , we get the local fitting energy term of the acoustic image level set energy function, which can be expressed as the following formula:

[0064] Among them, K is the Gaussian kernel function, is the local fitting energy term parameter, is the length penalty parameter, H() is the Heaviside function, which can be expressed as the following formula:

[0065] In order to maintain the signed distance property of the level set function during the evolution of the zero-order curve of the level set, the level set function Maintain smoothness, be more accurate when calculating derivatives, avoid time-consuming reinitialization operations, and introduce distance regularization terms in the level set energy function , which can be expressed as the following formula:

[0066] in, is the distance regularization parameter, is the potential function, which can be expressed as the following formula:

[0067] Among them, in order to make the segmentation contour of the acoustic target more accurate and smooth, and to stop the contour evolution iterative process at the correct segmentation position, the superpixel segmentation contour constraint term is introduced into the level set energy function to drive the zero-order segmentation curve of the level set close to the superpixel boundary, eliminate image noise interference, and obtain a more accurate target contour. The superpixel segmentation contour constraint term can be expressed as the following formula:

[0068] in, is the superpixel constraint parameter is the level set stop function. According to the principle of superpixel segmentation and level set constraint, the stop function can be expressed as the following formula:

[0069] in, is the gradient at pixel x, is the point set at the superpixel boundary, so that the zero-level segmentation curve stops evolving when it evolves near the superpixel boundary.

[0070] Step S302: obtaining a dual-region level set evolution equation based on the level set energy function.

[0071] Among them, the dual-region level set evolution equation can be obtained through the level set energy function, and the dual-region level set evolution equation can be expressed as the following formula:

[0072] in, is the divergence operation, It can be expressed as the following formula:

[0073] Step S303 , evolving the initial contour of the bright area and the initial contour of the dark area based on the dual-region level set evolution equation until the evolution converges to obtain the target contour.

[0074] For ease of understanding, Figure 6 A schematic diagram of contour evolution is provided in an embodiment of the present application, in which the initial contour of the bright area and the initial contour of the dark area are evolved by a dual-region level set evolution equation until the evolution converges, thereby obtaining the target contour.

[0075] Example 4 Corresponding to the above method embodiment, the embodiment of the present invention provides a device for detecting underwater silt distribution. Figure 7 A schematic diagram of the structure of a device for detecting underwater silt distribution provided by an embodiment of the present invention is shown in FIG. Figure 7 As shown, the underwater silt distribution detection device may include: The sonar data analysis module 401 is used to acquire sonar data and analyze it to obtain a sonar waterfall image.

[0076] The denoising processing module 402 is used to perform denoising processing on the sonar waterfall image based on the weight of the first pixel relative to the second pixel in the sonar waterfall image to obtain a target sonar waterfall image.

[0077] The super-pixel segmentation module 403 is used to perform super-pixel segmentation on the target sonar waterfall image based on a preset segmentation rule to obtain a super-pixel image.

[0078] The target region superpixel determination module 404 is used to analyze the local neighborhood mean and variance of the superpixel image to determine the target region superpixel.

[0079] The target initial contour determining module 405 is used to determine the target initial contour based on the target area superpixels.

[0080] The target contour determination module 406 is used to perform level set segmentation processing on the initial target contour to determine the target contour.

[0081] The underwater silt distribution detection device provided by the embodiment of the present invention can obtain sonar data and analyze it to obtain a sonar waterfall image, perform denoising on the sonar waterfall image based on the weight of the first pixel relative to the second pixel in the sonar waterfall image to obtain a target sonar waterfall image, perform superpixel segmentation on the target sonar waterfall image based on a preset segmentation rule to obtain a superpixel image, analyze the local neighborhood mean and variance of the superpixel image to determine the target area superpixel, determine the target initial contour based on the target area superpixel, and perform level set segmentation on the target initial contour to determine the target contour. In this method, the contour of underwater silt is determined by obtaining and analyzing sonar data and performing image denoising, superpixel segmentation and level set segmentation, thereby ensuring the speed and accuracy of contour determination.

[0082] In some embodiments, the denoising processing module is further used to perform denoising processing on the sonar waterfall image based on the weight of the first pixel relative to the second pixel in the sonar waterfall image by the following formula to obtain a target sonar waterfall image: ;in, ; represents the weight of the first pixel qi relative to the second pixel p; Z(p) represents the weight normalization factor, Represents the Euclidean distance between the second pixel p and the center of the first pixel qi.

[0083] In some embodiments, the superpixel segmentation module is also used to segment the target sonar waterfall image according to S×S pixels to obtain multiple superpixel areas; perform pixel seed point iteration and pixel label iteration on the center of each superpixel area; perform center clustering and local superpixel seed point iteration and pixel label iteration in a specified neighborhood of each superpixel area to determine the clustering distance; and allocate each pixel based on the clustering distance to obtain a superpixel image.

[0084] In some embodiments, the designated neighborhood is a 2S×2S neighborhood; the superpixel segmentation module is further used to determine the clustering distance by the following formula: ; Where D represents the clustering distance; d c is the pixel distance; d s is the pixel coordinate distance, indicating the position distance between the pixel point i at the coordinates (xi, yi) and the seed point j at the coordinates (xj, yj); m is a constant with a value of 1-40; dt represents the LDZP texture description value distance.

[0085] In some embodiments, ; where d c is the pixel distance, indicating that the pixel value is NL[v] i The pixel point i and the pixel value are NL[v] j The difference in pixel values ​​of seed point j; ;in, d s is the pixel coordinate distance, indicating the coordinate (x i ,y i ) i With coordinates ( x j , y j ) at the seed point j The location distance; ;in, d t LDZP texture description value distance, expressed in pixels i With pixels j The distance between the N-neighborhood LDZP texture description values.

[0086] In some embodiments, the mud area in the sonar waterfall image includes a background area, a bright area and a dark area, and the bright area and the dark area are adjacent to each other; the target area superpixel determination module is also used to determine a first relationship between the superpixel value of the superpixel in the background area and the local neighborhood mean and variance; determine a second relationship between the superpixel value of the target bright area superpixel in the bright area and the local area mean and variance; determine a third relationship between the superpixel value of the target dark area superpixel in the dark area and the local area mean and variance; determine a local statistical information satisfaction relationship of the target bright area based on the first relationship, the second relationship and the third relationship; determine the target area superpixel based on the local statistical information satisfaction relationship and pre-set data requirements.

[0087] In some embodiments, the target initial contour includes: an initial contour of a bright area corresponding to the target bright area and an initial contour of a dark area corresponding to the target dark area; a target contour determination module is further used to construct a level set energy function based on distance regularization for the target sonar waterfall image based on the initial contour of the bright area and the initial contour of the dark area; a dual-region level set evolution equation is obtained based on the level set energy function; based on the dual-region level set evolution equation, the initial contour of the bright area and the initial contour of the dark area are evolved until the evolution converges to obtain the target contour.

[0088] The device provided in the embodiment of the present invention has the same implementation principle and technical effects as those of the aforementioned method embodiment. For the sake of brief description, for matters not mentioned in the device embodiment, reference may be made to the corresponding contents in the aforementioned method embodiment.

[0089] Example 5 The embodiment of the present invention also provides an electronic device for executing the above-mentioned underwater silt distribution detection method; see Figure 8 A structural schematic diagram of an electronic device is shown, which includes a memory 500 and a processor 501, wherein the memory 500 is used to store one or more computer instructions, and the one or more computer instructions are executed by the processor 501 to implement the above-mentioned underwater silt distribution detection method.

[0090] Further, Figure 8 The electronic device shown further includes a bus 502 and a communication interface 503 , and the processor 501 , the communication interface 503 and the memory 500 are connected via the bus 502 .

[0091] The memory 500 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk storage. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 503 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used. The bus 502 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 8 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0092] The processor 501 can also be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 501. The above processor 501 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in conjunction with the embodiments of the present invention can be directly embodied as a hardware decoding processor for execution, or a combination of hardware and software modules in the decoding processor for execution. The software module may be located in a storage medium mature in the art, such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 500, and the processor 501 reads the information in the memory 500 and completes the steps of the method of the above embodiment in combination with its hardware.

[0093] An embodiment of the present invention also provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the above-mentioned underwater silt distribution detection method. The specific implementation can be found in the method embodiment, which will not be repeated here.

[0094] The computer program product of the underwater silt distribution detection method provided in the embodiment of the present invention includes a computer-readable storage medium storing a non-volatile program code executable by a processor. The instructions included in the program code can be used to execute the method described in the previous method embodiment. The specific implementation can be found in the method embodiment, which will not be repeated here.

[0095] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0096] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0097] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0098] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0099] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that can be executed by a processor. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., various media that can store program codes.

[0100] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for detecting underwater silt distribution, characterized in that: The method comprises: Acquire sonar data and parse it to obtain sonar waterfall images; De-noising the sonar waterfall image based on the weight of the first pixel relative to the second pixel in the sonar waterfall image to obtain a target sonar waterfall image; Performing super-pixel segmentation on the target sonar waterfall image based on a preset segmentation rule to obtain a super-pixel image; Analyzing the local neighborhood mean and variance of the superpixel image to determine the target area superpixel; Determine an initial contour of the target based on the superpixels of the target area; The target contour is determined by performing level set segmentation processing on the initial target contour.

2. The method according to claim 1, characterized in that The step of performing denoising on the sonar waterfall image based on the weight of the first pixel relative to the second pixel in the sonar waterfall image to obtain the target sonar waterfall image includes: The target sonar waterfall image is obtained by denoising the sonar waterfall image based on the weight of the first pixel relative to the second pixel in the sonar waterfall image using the following formula: ; in, ; Indicates the first pixel q i Relative to the second pixel p The weight of Z ( p ) represents the weight normalization factor, Represents the second pixel p With the first pixel q i Euclidean distance between centers.

3. The method according to claim 2, characterized in that The step of performing super-pixel segmentation on the target sonar waterfall image based on a preset segmentation rule to obtain a super-pixel image includes: Segmenting the target sonar waterfall image into S×S pixels to obtain a plurality of super-pixel regions; Perform pixel seed point iteration and pixel label iteration on the center of each superpixel region; Performing center clustering and local superpixel seed point iteration and pixel label iteration in a specified neighborhood of each superpixel region to determine a clustering distance; Each pixel is assigned based on the cluster distance to obtain a superpixel image.

4. The method according to claim 3, characterized in that The designated neighborhood is a 2S×2S neighborhood; performing center clustering and performing local superpixel seed point iteration and pixel label iteration in the designated neighborhood of each superpixel region to determine the clustering distance includes: The cluster distance is determined by the following formula: ; Where D represents the cluster distance; d c is the pixel distance; d s is the pixel coordinate distance, indicating the coordinate ( x i , y i ) and the pixel point i at the coordinate ( x j , y j ) is the position distance of the seed point j; m is a constant; d t Indicates the LDZP texture description value distance.

5. The method according to claim 4, characterized in that ;in, d c is the pixel distance, indicating that the pixel value is NL[v] i Pixels i and the pixel value is NL[v] j The seed point j The difference in pixel values; ;in, d s is the pixel coordinate distance, indicating the coordinate ( x i , y i ) i With coordinates ( x j , y j ) at the position distance of the seed point j; ;in, d t is the LDZP texture description value distance, which represents the N-neighborhood LDZP texture description value distance between pixel i and pixel j.

6. The method according to claim 5, characterized in that The silt area in the sonar waterfall image includes a background area, a bright area and a dark area, and the bright area and the dark area are adjacent to each other; the local neighborhood mean and variance of the superpixel image are analyzed to determine the superpixel of the target area, including: Determining a first relationship between a superpixel value and a local neighborhood mean and variance of a superpixel in the background region; Determine a second relationship between a superpixel value of a target bright area superpixel in the bright area and a local area mean and variance; Determining a third relationship between a superpixel value of a target dark area superpixel in the dark area and a local area mean and variance; Determining based on the first relationship, the second relationship and the third relationship that the local statistical information of the target bright area satisfies a relationship; The target area superpixel is determined based on the local statistical information satisfying the relationship and the pre-set data requirements.

7. The method according to claim 6, characterized in that The target initial contour includes: a bright area initial contour corresponding to the target bright area and a dark area initial contour corresponding to the target dark area; performing level set segmentation processing on the target initial contour to determine the target contour includes: Constructing a level set energy function based on distance regularization for the target sonar waterfall image based on the initial contour of the bright area and the initial contour of the dark area; Based on the level set energy function, a dual-region level set evolution equation is obtained; The bright area initial contour and the dark area initial contour are evolved based on the dual-region level set evolution equation until the evolution converges to obtain the target contour.

8. A device for detecting underwater silt distribution, characterized in that: The device comprises: Sonar data analysis module, used to obtain sonar data and analyze it to obtain sonar waterfall images; A denoising processing module, configured to perform denoising processing on the sonar waterfall image based on a weight of a first pixel relative to a second pixel in the sonar waterfall image to obtain a target sonar waterfall image; A superpixel segmentation module, used for performing superpixel segmentation on the target sonar waterfall image based on a preset segmentation rule to obtain a superpixel image; A target area superpixel determination module is used to analyze the local neighborhood mean and variance of the superpixel image to determine the target area superpixel; A target initial contour determination module, used to determine the target initial contour based on the target area superpixels; The target contour determination module is used to perform level set segmentation processing on the initial target contour to determine the target contour.

9. An electronic device, characterized in that: It comprises a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the underwater silt distribution detection method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions. When the computer-executable instructions are called and executed by the processor, the computer-executable instructions prompt the processor to implement the underwater silt distribution detection method according to any one of claims 1 to 7.

Images